A major reason why researchers may want to band together outside a university is to share fixed costs, particularly technological ones.
Academic code is famously bad, because it is not written by professional software engineers, but grad students and postdocs whose primary job is publishing papers. A lot of research fields (biology is the one I’m most familiar with) can see large potential gains from high-quality software engineering that aren’t easily captured in universities because it’s hard for an individual professor to get an engineer on his grant.
The Broad Institute has done a lot of this already—they have a large team of engineers in addition to their research scientists, and so they can build a lot of shared infrastructure for computational biology projects. In my opinion this is very successful. When you see a large, well-maintained biological database available to the public, there’s a good chance it came from the Broad.
There are other shareable fixed costs associated with R&D, if you’re doing translational research. If you want to spin out actual drug candidates from drug development research, for instance, it can help to have business development, legal, and regulatory staff on retainer for the use of any scientist who needs their services. Universities aren’t always good at this, and take a hefty cut of the profits from any IP. The Chan Zuckerberg Initiative has this as a model—they want to enable translational research by partnering with biologists and having experienced people to help them with the tech stuff and the business stuff.
Why would a top researcher want to join an org like this? For the same reason that Geoffrey Hinton would go to Google—big resources (such as huge amounts of computational power) that he can’t get at a university. The “syndicate” format might be attractive to people who want university-style autonomy (which you can’t usually get in industry) while still having access to business-style resources. Though there’s probably an inherent tradeoff between “the scientists run the org without being micromanaged” and “the org continues to bring in enough money to provide great common resources.”
Thanks for the pointer to the Broad Institute, it does look like an interesting model.
Though there’s probably an inherent tradeoff between “the scientists run the org without being micromanaged” and “the org continues to bring in enough money to provide great common resources.”
I think it would need agreements between people so that a certain percentage of time was spent on sure fire things and could be somewhat managed, but give people enough time to work on the risky/interesting things. I’m interested in what Benquo learns from the communties he is is interested in, I hope there are lessons for this kind of endeavour.
A major reason why researchers may want to band together outside a university is to share fixed costs, particularly technological ones.
Academic code is famously bad, because it is not written by professional software engineers, but grad students and postdocs whose primary job is publishing papers. A lot of research fields (biology is the one I’m most familiar with) can see large potential gains from high-quality software engineering that aren’t easily captured in universities because it’s hard for an individual professor to get an engineer on his grant.
The Broad Institute has done a lot of this already—they have a large team of engineers in addition to their research scientists, and so they can build a lot of shared infrastructure for computational biology projects. In my opinion this is very successful. When you see a large, well-maintained biological database available to the public, there’s a good chance it came from the Broad.
There are other shareable fixed costs associated with R&D, if you’re doing translational research. If you want to spin out actual drug candidates from drug development research, for instance, it can help to have business development, legal, and regulatory staff on retainer for the use of any scientist who needs their services. Universities aren’t always good at this, and take a hefty cut of the profits from any IP. The Chan Zuckerberg Initiative has this as a model—they want to enable translational research by partnering with biologists and having experienced people to help them with the tech stuff and the business stuff.
Why would a top researcher want to join an org like this? For the same reason that Geoffrey Hinton would go to Google—big resources (such as huge amounts of computational power) that he can’t get at a university. The “syndicate” format might be attractive to people who want university-style autonomy (which you can’t usually get in industry) while still having access to business-style resources. Though there’s probably an inherent tradeoff between “the scientists run the org without being micromanaged” and “the org continues to bring in enough money to provide great common resources.”
Thanks for the pointer to the Broad Institute, it does look like an interesting model.
I think it would need agreements between people so that a certain percentage of time was spent on sure fire things and could be somewhat managed, but give people enough time to work on the risky/interesting things. I’m interested in what Benquo learns from the communties he is is interested in, I hope there are lessons for this kind of endeavour.